Manager 23March11

From GO Wiki
Jump to: navigation, search

Agenda - carried over from March 9th

Attending: Rama, Pascale, Suzi, Judy, David, Emily, PaulT
Minutes by Rama


  • Can everyone check if there are items from the previous meeting we need to carry over? Manager_23Feb11

Old Business

  • MGI Curator's report on changes made to term definitions for annotators to refine their annotations if they think it is warranted. Amelia has written the report; http://www.geneontology.org/scratch/def_diffs.html. his data could be updated once a week.

ACTION: format check, then send out to the GO list. This data could be updated once a week. Location in the scratch directory correct, or should be in a directory for reports?Need to talk with Chris.
Discussion

  • There should a new directory called QC and these reports should go into that directory once a week. Each group should be checking these reports every week.
  • There are lot of things to follow up for curators. How should curators/groups be reminded about these tasks?
  • May be each group should have a representative to check these reports and the MF-BP inferences, PAINT etc.
  • CVS can send an email out when files are checked in. For now may be CVS should send emails out to remind groups?
  • Bring this up in the GOC meeting to find out how groups would like to be reminded
  • There was a problem with generating the PAINT GAF files. Suzi will check with MikeC.

Discussion

  • Emily will send around the URl for suggestions.
  • Everybody should look at the GOC grant to see if anything needs further discussion (Aims/Deliverable)
  • Moving forward, quarterly reports (audits) will be due to assess progress.
  • We need a tracking system for software projects. PIs will talk about this.
  • RT system at EBI-it is an open source system and we can try it for gohelp and may be for project tracking. We can try it and then decide.

New Discussion Topics

Ontology development

(Pascale): Obsoleting the xx-binding terms in the transcription overhaul: Can this be put on hold until we have a GO-wide strategy?

Discussion

  • It seems like we are creating binding terms and also obsoleting some..thers is some inconsistency.
  • The main reason for obsoleting these promoter binding terms is because there is no good definition for what a promoter is and we can't manage all the promoters.
  • We could use SO to represent these features in a structured way and we will clean up these annotations when we do ref.genome curation
  • we should follow a 'response_to' approach
  • This led to a discussion on post and pre-composing terms and use of col-16. This will be an agenda in the GOC meeting (it is also one of the aims). Annotation team should think about this issue and should have a plan.
  • People are not populating col-16. Groups will start doing it only if they are forced to. The other problem is none of the downstream tools can accommodate this new data. Groups don't have software support to show col-16. GOC has to come with a set of things that the groups should absolutely get done and give some time frame. AmiGO is yet to accommodate col-16. Perhaps that should be the first step?

Items from the Annotation and Reference genome Groups

  • New Reference Genome targets? (as the apoptosis set is to be frozen until after the June Content meeting)

Discussion: Transcription in heart development will be worked on instead.

Discussion: Good idea. We could also say this should we write a paper on GAF.

Discussion

  • all looks good. Chris mentioned that he will demo an interface where one can upload a GAF and find out all the errors.
  • Technical priorities for Annotation Groups. Not Discussed- will be carried over to next week.

Would it be desirable to generate a page for Technical Priorities for Annotation Groups that we require of annotation groups (e.g. development of annotation file formats, incorporating PAINT and m2p inferences, adding column 16 data etc.)that often involves software engineer time. At GOC meetings many ideas are presented, some are actioned but are not followed up by MODs. Its understandable that some groups are reluctant to use valuable programmer time until they have been convinced that the GOC is determined to go in a certain direction. The page could summarize the top jobs, with links to final specifications of any file formats.

  • Tracking gene product annotation status form each database

using the gp annotation and information format

Reminder: this was an item brought up at the GO Consortium meeting in Stanford March 2010. Here are the minutes from the discussion - One of the action items was: We will roll out new split GAF for all groups

  • Proposal for definition of Comprehensively annotated

- this tag is attached at the level of the gene product and must be accompanied by a timestamp.

Definition:

Indicates that a gene product has been the focus of a manual annotation effort whereby a curator has reviewed the current literature and has annotated to the principal functions, process and component GO terms. It is not required for a curator should have annotated every single paper published on a gene product if additional annotations would only duplicate information currently provided.

The timestamp is an essential component of this label, as it indicates the date at which the curator last reviewed the gene product's annotation set and literature available.

It is possible that a gene product may have papers that were published after the 'comprehensively annotated' timestamp that would be suitable for GO annotation but that have not yet been added added.

Similarly, GO annotations to a comprehensively annotated gene product with a later timestamp may exist which have been created during routine annotation update procedures, or as a by-product from a annotation efforts focusing on a different gene product.

Each time a curator re-reviews the entire annotation set for a previously 'comprehensively annotated'-tagged gene product, the associated timestamp should be updated.